|Publication number||US7774339 B2|
|Application number||US 11/811,596|
|Publication date||Aug 10, 2010|
|Priority date||Jun 11, 2007|
|Also published as||US20080306937, WO2008154197A1|
|Publication number||11811596, 811596, US 7774339 B2, US 7774339B2, US-B2-7774339, US7774339 B2, US7774339B2|
|Inventors||Ryen W. White, Mikhail Bilenko, Nicholas E. Craswell, Michael M. Cameron, Hugh E. Williams|
|Original Assignee||Microsoft Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (15), Non-Patent Citations (5), Referenced by (23), Classifications (6), Legal Events (5)|
|External Links: USPTO, USPTO Assignment, Espacenet|
Information retrieval is a process by which a user submits a query to a search engine, and the search engine launches the query against content (or an index of that content) to identify content that is responsive, or otherwise relevant to, the query input by the user. The results returned by the search engine often include hyperlinks to the identified content, in rank order, based on some relevance measure. The results may also include some descriptive material associated with the hyperlink, such as a few phrases from the content.
In an effort to improve information retrieval systems, some work has been done involving alternative query formulations, known as “query suggestions”, which are offered to a user following an initial query. The query suggestion allows the user to modify the specification of their informational needs provided to the search engine, which often leads to improved retrieval performance. Some current query suggestion approaches draw upon the query reformulation behavior of other users, to make query recommendations based on previous user interactions with the search engine.
However, this type of interaction-based approach to query suggestion has its limitations. It can be less effective in assisting a user when the information needed by the user is exploratory in nature. This is because a large portion of user activity for such information needs may occur beyond search engine interactions. For instance, in cases where directed searching is only a fraction of users' information-seeking behavior, the utility of users' clicks over the space of top-ranked results may be limited in its effectiveness, because it does not cover the subsequent browsing behavior of those users.
Some browser plug-ins and proxy server logs provide access to the browsing patterns of users that transcend search engine interactions. These data have been used to improve search result ranking, but only did so by considering page visitation statistics independently of one another. It did not take into account the pages' relative positions on post-query browsing paths. Other systems have used this type of information to improve retrieval accuracy by using sequences of consecutive query reformulations. However, this type of system does not consider a user's interactions beyond the reformulation of queries on the search result page.
One system has investigated the possibility of supporting navigation based on the browse trail of other users within a specific domain. This system operated by finding the most popular pages within a given domain. This system simply offers several within-domain shortcuts below the title of a returned search result. The shortcuts appear to be sub-links from a homepage listing, to a set of pages that are directly accessible from that home page listing.
The discussion above is merely provided for general background information and is not intended to be used as an aid in determining the scope of the claimed subject matter.
It has been found that user navigation that follows search engine interactions provides implicit endorsement of resources (such as web resources) that are preferred by users, and which may be particularly valuable for exploratory search tasks. Thus, a combination of past searching and browsing user behavior is analyzed to identify additional information that augments search results delivered by a search engine. The additional information may include a display of hyperlinks to locations which are derived from the past searching and browsing user behavior, given a specific input query. The additional information may be provided to supplement web search results.
This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter. The claimed subject matter is not limited to implementations that solve any or all disadvantages noted in the background.
Search engine 104 illustratively receives a query 118 from user 102, and searches index 104. The query 118 is also illustratively provided (either through search engine 104, or directly from the user through an appropriate user interface) to search trail processor 108. Search engine 104 illustratively returns organic search results (such as a listing of hyperlinks to documents containing information deemed relevant to query 118), and search trail processor 108 illustratively returns additional sites/domains 120, based not only on query 118, but based on processing of query/interaction logs 122.
In any case, search trail processor 108 illustratively accesses query/interaction logs 122. This is indicated by block 150 in
Search trail processor 108 thus reconstructs the temporally ordered sequences of viewed pages (i.e., the trails) from the query/interaction logs 122. This is indicated by block 152 in
In order to identify the popular destinations, starting with every query submission to a search engine, processor 108 follows the corresponding trail to a point of termination, where it is assumed that a user has completed the information-seeking activity. In one embodiment, the trails include pages that are search result pages, search engine home pages, or pages connected to a search result page through a sequence of clicked hyperlinks. Extracting search trails in this way helps to eliminate ambiguity associated with multitasking, where users run multiple searches concurrently. Since users may open a new browser window for each task, each task has its own browser trail because it is associated with a new browser instance, and it also has a corresponding distinct search trail.
Search trail processor 108 terminates the search trail when a user returns to the user's home page, checks electronic mail, logs into an online service, types a URL, or visits a bookmarked page. Processor 108 also terminates the search trail if a page is viewed by the user for more than thirty minutes with no activity, or where the user closes the active browser window. If a page in the search trail meets any of these criteria, the trail is assumed to terminate on the previous page. Of course, additional or different criteria could be used.
As shown again with respect to
Search trail processor 108 then generates additional information from the session trails 131 and query trails 133. Again, the following information can be generated by search trail processor 108 by pre-computing it, by pre-computing only part of it, or by computing it all during runtime.
In any case, search trail processor 108 may illustratively identify “destinations”, “interactions hubs”, and “way stations” or other items, from search trails 110. These sites or domains can be provided as additional information 120, along with the organic search results, to assist the user.
Destinations are illustratively web pages or web domains where other users have ended up, following submission of a query (and potentially after traversing multiple queries in-between the initial query and the ultimate destination). Following visits to destinations, a user terminates the search.
Interaction hubs are web pages or web domains that other users interact intensively with following submission of a query. This interaction is typically characterized by viewing pages that the hubs link to, and then returning to the hub and viewing more pages linked from it. Users tend to find some utility in such locations.
Way stations are illustratively web pages or domains that other users pass through on route to other pages or domains. Way stations may contain little or no relevant information to the query, but they may well be required to get to pages that contain such information in an efficient manner. Identifying destinations, interaction hubs and way stations, among other things, is indicated by block 154 in
Search trail processor 108 illustratively identifies trails 130 or 132 that are associated with experts. This is indicated by optional block 156 in
In another embodiment, a user may be identified as having domain expertise based on how often this particular user looks at information on a particular topic. For example, a user that frequently submits many queries on “physics”, “astronomy”, and “space” is likely to be more informed about the best places to visit for searches in that subject matter area than an average person.
The interaction patterns of all users searching for a given topic, and identified by interaction logs 122, may overwhelm the interactions performed by experts in the area. Therefore, search trail processor 108, in returning the additional sites 120, may weight those visited by experts more heavily than others. Of course, other filtering techniques can be used to filter the destinations based on user expertise so that the user 102 can see which destinations are popular and useful to people with domain expertise.
Search trail processor 108 may also identify a set of suggested sites or domains in temporal order, that can be provided to the user. This is indicated by block 158 in
There are a wide variety of ways which the user can be walked through such an ordered set of pages or domains. For instance, following submission of the query by the user, and after receiving an indication that the user desires to follow the ordered set of pages, the user can be provided with a combo box that displays the ordered pages or domains, and also provides navigation options. This may be provided as a plug-in component, or otherwise. Similarly, the user can be provided with a frame launchable from the search result page that offers similar functionality without the need to have a toolbar installed. In any case, the set of ordered sites or domains can be identified by search trail processor 108.
It may also happen that a user mis-types the query. Some current approaches to correcting spelling mistakes in a query involve the presentation of query refinements, usually in the form “did you mean [blank]” where a reformulated query is inserted in place of [blank]. In such current systems, the user is then required to click on the reformulated query, and that query is then launched against the index or web content. This type of system has a number of problems. For instance, if the query that the user actually submitted is a valid query, then the spelling correction system will not be deployed, and the user will be presented with a list of seemingly irrelevant results (although they may be relevant to the actual query submitted by the user).
In contrast, in one embodiment, system 100 does not require the user to discover that the query has been mistyped. Instead, by processor 108 processing the query/interaction logs 122, processor 108 can identify the pages or domains that others have visited following the subsequent correction of the query. For instance, processor 108 may determine that, given the query, the user is likely to reformulate it to another query, and then to visit a page or domain returned as a result from launching the reformulated query. Instead of requiring the user to go through all of these steps, processor 108 simply identifies the sites or domains responsive to the anticipated query reformulation and returns those to the user automatically. This is indicated by block 160 in
An example may be helpful. Assume that the input query is “acys”. This happens to be a valid query, in that it signifies a term which is sometimes used to search for a site, but the site is not very popular. In fact, most users refine that query to “macys” and visit “macys.com”. Instead of asking the user to confirm that the user meant to type “macys”, processor 108 simply presents the additional sites/domains provided along with the enhanced results 119 “macys.com” as well as some indication that the input query has been reformulated to “macys” to yield this result. This has great potential to speed up user searching and improve user satisfaction. It will be noted that processor 108 does not require a user to relaunch a reformulated query, but simply anticipates that the user, if given the opportunity, would in fact reformulate the query and launch it. Therefore, processor 108 simply returns the results to the anticipated reformulation of the query, saving the user time.
It should be noted that all of these items of information generated by search trail processor 108 can be computed beforehand, or during runtime once the user has input a runtime query, or a combination of both.
Search trail processor 108 then estimates additional sites or domains, in addition to the search results generated by search engine 104, and possibly filters them based on the expertise information described above with respect to block 156 in
Estimation of these sites and domains can be performed in a wide variety of different ways. In one illustrative embodiment, a term-based site/domain prediction model is used. In accordance with the exemplary discussion herein, a prediction model for predicting destinations 130 will be described, although a similar model could be generated for interaction hubs 132 and way stations 134 as well.
As discussed above, two types of destinations are first extracted. One is a destination for a query trail, and the other is a destination for a session trail. These are referred to as query destinations and session destinations, respectively. For both destination types, a corpus of query-destination pairs is obtained and used to construct term-vector representations of destinations that are analogous to the known tf.idf (term frequency.inverse document frequency) document representation in conventional information retrieval systems. Then, given a new query q consisting of k terms t1 . . . tk, we identify highest-scoring destinations using the following similarity function:
where query and destination term weights, wq(ti) and wd(ti), are computed using standard tf.idf weighting and query and user session-normalized smoothed tf.idf weighting, respectively.
Of course, other scoring functions could be used as well. In any case, given the query, the destinations 210, interaction hubs 212 and way stations 214 are estimated. Processor 108 and search engine 104, then return the organic search results enhanced with the additional sites or domains 120 (collectively referred to as 121 in
The organic search results, enhanced with the additional sites or domains, can be used in a plurality of different ways, three examples of which are shown in
For instance, assume that a user has input a query “dogs” but ends up navigating through a series of web pages to a destination “petfoods.com”. The page which the user eventually settled on may not even be topically relevant to the query. The presentation of such popular locations provides a mechanism through which users can be taken to places of interest (web pages or domains of interest) to other users, potentially bypassing pages that lie in between. Navigating all the pages in between may be a time consuming burden to the user. Therefore, the presentation of such pages can effectively transport the user several steps into the future along the query trail (or session trail) identified by processor 108. This can save users time in terms of subsequent query reformulations or browsing through pages to find desired hyperlinks.
In accordance with one embodiment, a user can control how far into the future they are transported along the prior query trail (or session trail) identified by search trail processor 108. For instance, the user may be presented with links to pages that were navigated to by prior users, immediately before the prior users launched a next query. Similarly, the user may be presented with pages that were navigated to from a next search result or that were navigated to immediately before the end of a prior search session in the session trail identified by processor 108. In any case, presenting the user with page or domain destinations that were navigated to by prior users who entered a similar query is indicated by presenting the user with teleportation results 218 in
Search trail processor 108 can also present the user with results in response to anticipated query reformulation. This is indicated by block 220 in
A third presentation of enhanced results to the user is indicated by block 222 in
It will be noted of course that processor 108 can provide a wide variety of different suggested additional sites or domains to the user, and those listed with respect to
Once the organic search results, enhanced with the additional sites/domains 121 are presented to user 102 (as indicated by block 216 in
Based on the user's browse path, search trail processor 108 can present different (or the same) suggested additional sites/domains to the user after the user leaves the result page. This is indicated by block 228 in
In any case, in one embodiment, the additional sites/domains 120, such as destinations 130, interaction hubs 132 and way stations 134 (whether they be the original additional sites/domains 120 presented as part of results 121, or dynamically changing additional sites/domains 120) can be offered to the user in a small semi-transparent overlay on the page then being viewed by the user. It may be desirable to provide this overlay on the top right-hand corner of the web page, for instance, because that is a location that is typically devoid of substantive page content. However, other places on the display could just as easily be used. The additional sites/domains 120 presented to the user in this way may illustratively point users to places (web sites or domains) where others who visited the domain currently being visited by the current user, ended up, interacted with extensively, or simply passed through. Of course, it will be noted that modifying the sites/domains and presenting them to the user even after the user leaves the results page need not be performed.
Embodiments are operational with numerous other general purpose or special purpose computing system environments or configurations. Examples of well-known computing systems, environments, and/or configurations that may be suitable for use with various embodiments include, but are not limited to, personal computers, server computers, hand-held or laptop devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, telephony systems, distributed computing environments that include any of the above systems or devices, and the like.
Embodiments may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Some embodiments are designed to be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules are located in both local and remote computer storage media including memory storage devices.
With reference to
Computer 310 typically includes a variety of computer readable media. Computer readable media can be any available media that can be accessed by computer 310 and includes both volatile and nonvolatile media, removable and non-removable media. By way of example, and not limitation, computer readable media may comprise computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by computer 110. Communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared and other wireless media. Combinations of any of the above should also be included within the scope of computer readable media.
The system memory 330 includes computer storage media in the form of volatile and/or nonvolatile memory such as read only memory (ROM) 331 and random access memory (RAM) 332. A basic input/output system 333 (BIOS), containing the basic routines that help to transfer information between elements within computer 310, such as during start-up, is typically stored in ROM 331. RAM 332 typically contains data and/or program modules that are immediately accessible to and/or presently being operated on by processing unit 320. By way of example, and not limitation,
The computer 310 may also include other removable/non-removable volatile/nonvolatile computer storage media. By way of example only,
The drives and their associated computer storage media discussed above and illustrated in
A user may enter commands and information into the computer 310 through input devices such as a keyboard 362, a microphone 363, and a pointing device 361, such as a mouse, trackball or touch pad. Other input devices (not shown) may include a joystick, game pad, satellite dish, scanner, or the like. These and other input devices are often connected to the processing unit 320 through a user input interface 360 that is coupled to the system bus, but may be connected by other interface and bus structures, such as a parallel port, game port or a universal serial bus (USB). A monitor 391 or other type of display device is also connected to the system bus 321 via an interface, such as a video interface 390. In addition to the monitor, computers may also include other peripheral output devices such as speakers 397 and printer 396, which may be connected through an output peripheral interface 395.
The computer 310 is operated in a networked environment using logical connections to one or more remote computers, such as a remote computer 380. The remote computer 380 may be a personal computer, a hand-held device, a server, a router, a network PC, a peer device or other common network node, and typically includes many or all of the elements described above relative to the computer 310. The logical connections depicted in
When used in a LAN networking environment, the computer 310 is connected to the LAN 371 through a network interface or adapter 370. When used in a WAN networking environment, the computer 310 typically includes a modem 372 or other means for establishing communications over the WAN 373, such as the Internet. The modem 372, which may be internal or external, may be connected to the system bus 321 via the user input interface 360, or other appropriate mechanism. In a networked environment, program modules depicted relative to the computer 310, or portions thereof, may be stored in the remote memory storage device. By way of example, and not limitation,
Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US6006225||Sep 1, 1998||Dec 21, 1999||Amazon.Com||Refining search queries by the suggestion of correlated terms from prior searches|
|US6411950||Nov 30, 1998||Jun 25, 2002||Compaq Information Technologies Group, Lp||Dynamic query expansion|
|US6671681||May 31, 2000||Dec 30, 2003||International Business Machines Corporation||System and technique for suggesting alternate query expressions based on prior user selections and their query strings|
|US6772150||Mar 22, 2000||Aug 3, 2004||Amazon.Com, Inc.||Search query refinement using related search phrases|
|US7062475||Nov 17, 2003||Jun 13, 2006||Alberti Anemometer Llc||Personalized multi-service computer environment|
|US7092936||Aug 22, 2001||Aug 15, 2006||Oracle International Corporation||System and method for search and recommendation based on usage mining|
|US7136845||Jul 12, 2001||Nov 14, 2006||Microsoft Corporation||System and method for query refinement to enable improved searching based on identifying and utilizing popular concepts related to users' queries|
|US7225407||Jun 28, 2002||May 29, 2007||Microsoft Corporation||Resource browser sessions search|
|US20050065774||Sep 20, 2003||Mar 24, 2005||International Business Machines Corporation||Method of self enhancement of search results through analysis of system logs|
|US20050177569||Apr 15, 2005||Aug 11, 2005||Dwayne Bowman||Identifying the items most relevant to a current query based on items selected in connection with similar queries|
|US20050278317||May 16, 2005||Dec 15, 2005||William Gross||Personalized search engine|
|US20060288000||Dec 30, 2005||Dec 21, 2006||Raghav Gupta||System to generate related search queries|
|US20070073641||Sep 23, 2005||Mar 29, 2007||Redcarpet, Inc.||Method and system for improving search results|
|US20090030876 *||Jan 23, 2004||Jan 29, 2009||Nigel Hamilton||Method and system for recording search trails across one or more search engines in a communications network|
|WO2005069161A1||Dec 15, 2004||Jul 28, 2005||Nigel Hamilton||Method and system for recording search trails across one or more search engines in a communications network|
|1||Christopher Holscher et al., "Web Search Behavior of Internet Experts and Newbies" http://www.9.org/w9cdrom/81/81.html pp. 1-16 , Jun. 2000.|
|2||Eugene Agichtein et al., "Learning User Interaction Models for Predicting Web Search Result Peferences" 2006, 8 pages.|
|3||Lori Lorigo et al., "Information Processing and Management: an International Journal" vol. 42, Issue 4, Jul. 2006, pp. 1123-1131 ISSN: 0306-4573.|
|4||PCT International Search Report Application No. PCT/US2008/065477, Dated Jun. 2, 2008.|
|5||Yumiko Seki et al., "User's Behavior and URL Analysis at EC Sites", pp. 87-92, 1999.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8122021 *||Oct 6, 2008||Feb 21, 2012||Microsoft Corporation||Domain expertise determination|
|US8402024 *||Jan 16, 2012||Mar 19, 2013||Microsoft Corporation||Domain expertise determination|
|US8429184||Apr 23, 2013||Collarity Inc.||Generation of refinement terms for search queries|
|US8438178||Jun 25, 2009||May 7, 2013||Collarity Inc.||Interactions among online digital identities|
|US8442972||Oct 11, 2007||May 14, 2013||Collarity, Inc.||Negative associations for search results ranking and refinement|
|US8650172 *||Mar 1, 2010||Feb 11, 2014||Microsoft Corporation||Searchable web site discovery and recommendation|
|US8812541||Mar 12, 2013||Aug 19, 2014||Collarity, Inc.||Generation of refinement terms for search queries|
|US8875038||Jan 19, 2011||Oct 28, 2014||Collarity, Inc.||Anchoring for content synchronization|
|US8903810||Oct 16, 2008||Dec 2, 2014||Collarity, Inc.||Techniques for ranking search results|
|US8930357||Nov 6, 2012||Jan 6, 2015||Microsoft Corporation||Domain expertise determination|
|US8972397||Mar 11, 2010||Mar 3, 2015||Microsoft Corporation||Auto-detection of historical search context|
|US8983996 *||Oct 31, 2011||Mar 17, 2015||Yahoo! Inc.||Assisted searching|
|US9268864 *||Nov 21, 2014||Feb 23, 2016||Microsoft Technology Licensing, Llc||Domain expertise determination|
|US20090119261 *||Oct 16, 2008||May 7, 2009||Collarity, Inc.||Techniques for ranking search results|
|US20090248661 *||Mar 28, 2008||Oct 1, 2009||Microsoft Corporation||Identifying relevant information sources from user activity|
|US20100088331 *||Oct 6, 2008||Apr 8, 2010||Microsoft Corporation||Domain Expertise Determination|
|US20110145226 *||Jun 16, 2011||Microsoft Corporation||Product similarity measure|
|US20110213761 *||Sep 1, 2011||Microsoft Corporation||Searchable web site discovery and recommendation|
|US20110225192 *||Mar 11, 2010||Sep 15, 2011||Imig Scott K||Auto-detection of historical search context|
|US20120117061 *||May 10, 2012||Microsoft Corporation||Domain expertise determination|
|US20130110863 *||Oct 31, 2011||May 2, 2013||Yahoo! Inc.||Assisted searching|
|US20140067783 *||Sep 6, 2012||Mar 6, 2014||Microsoft Corporation||Identifying dissatisfaction segments in connection with improving search engine performance|
|US20150081661 *||Nov 21, 2014||Mar 19, 2015||Microsoft Corporation||Domain expertise determination|
|U.S. Classification||707/722, 707/707, 707/706|
|Jul 16, 2007||AS||Assignment|
Owner name: MICROSOFT CORPORATION, WASHINGTON
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:WHITE, RYEN W.;BILENKO, MIKHAIL;CRASWELL, NICHOLAS E.;AND OTHERS;REEL/FRAME:019561/0608;SIGNING DATES FROM 20070607 TO 20070611
Owner name: MICROSOFT CORPORATION, WASHINGTON
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:WHITE, RYEN W.;BILENKO, MIKHAIL;CRASWELL, NICHOLAS E.;AND OTHERS;SIGNING DATES FROM 20070607 TO 20070611;REEL/FRAME:019561/0608
|Mar 21, 2014||REMI||Maintenance fee reminder mailed|
|Aug 10, 2014||LAPS||Lapse for failure to pay maintenance fees|
|Sep 30, 2014||FP||Expired due to failure to pay maintenance fee|
Effective date: 20140810
|Dec 9, 2014||AS||Assignment|
Owner name: MICROSOFT TECHNOLOGY LICENSING, LLC, WASHINGTON
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:MICROSOFT CORPORATION;REEL/FRAME:034542/0001
Effective date: 20141014